A Novel Bottom Up Approach to Bounding Potential Human Cancer Risks from Endogenous Chemicals Thomas...

A Novel Bottom Up Approach to Bounding Potential Human Cancer Risks

from Endogenous Chemicals

Thomas B. Starr, PhDTBS Associates, Raleigh NC

SOT RASS Webinar 13 November 2013

1

Typical Top Down Approach

• Cancer and exogenous exposure data extracted from epidemiology studies or laboratory animal bioassays

• Empirical or biologically-based dose-response models fit to cancer data vs exogenous exposure, e.g., airborne concentration, cumulative exposure

• Estimated BMDLx used to calculate upper bound unit risk for use in linear extrapolation or, alternatively, to compute MOEs for substances with nonlinear MOAs

2

The Bottom Up Approach

• Suitable for chemicals present in the body as a result of normal endogenous processes, e.g., metabolism

• Attributes all background risk P0 to tissue-specific endogenous background exposure C0

• Assumes linear dose-response for added risk AR vs exogenous exposure Cxss, with upper 95% confidence bound slope estimate (P0/C0)U: AR = (P0/C0)U · Cxss

• P0 from US SEER cancer statistics or bioassay dataC0 and Cxss data from short-term human/animal studies

3

Bottom Up Approach Elements

4

Bottom Up Approach Features

• Bounds low-dose cancer risk without using high dose cancer data from epidemiology studies or animal bioassays

• Provides an independent “reality check” on extrapolations from high-dose data

• Conservative:All background risk attributed to endogenous background exposureAssumes linearity at low dosesUpper bound estimates of lifetime risk

5

Estimating Steady-State Exogenous Adductsfrom Time Point-Specific Data

• Used one compartment model with constant forcing and firstorder elimination with half-life T1/2 = T · ln(2)

• For N2-hydroxymethyl-dG adducts in rats (10 ppm for 6 hrs)T1/2 = 63 hrs , T = 90.9 hrs (Swenberg 2012)

• At the end of one 6 hour exposure:Cxss = Cx6 /(1 – exp(-6/T)) = 15.65 · Cx6

• After two 6 hour exposures on consecutive days:Cxss = Cx30 /{[1 – exp(-6/T)] · [1 + exp(-24/T)]} = 8.85 · Cx30

6

One Compartment Model: Adduct Time Profile

7

Cx6

Cx30

Cxss

N2-Hydroxymethyl-dG Elimination Half-Life Data

One 6 hr exposure of rats to 10 ppm, Swenberg et al., 2012 8

New N2-Hydroxymethyl-dG Elimination Half-Life Data

Swenberg (unpublished data) 9

N2-hydroxymethyl-dG Adducts in Monkeys Exposed Twice for 6 hrs to 2 ppm 13CD2O

10

C0L

C0L

Cxss

Cxss

Cx30

Cx30

Comparison of Bottom Up and Top Down Upper Bound Added Risk Estimates

For NPC, ARBU = (3.44 x 10-4 · 2.21) / 2 = 0.038 x 10-2

= 29.8-fold lower than AREPA

For LEU, ARBU = (< 8.5 x 10-4 · 0.00912) / 2 = < 3.9 x 10-6

= > 14,615-fold lower than AREPA

11

Bottom Up Uncertainties (Human Analysis)

- P0 very precise due to large number of cases in US population of more than 300,000,000:

Annually, > 2,550 NPC, > 45,880 LEU

NPC P0 = 7.2500 x 10-4, P0U = 7.2656 x 10-4

LEU P0 = 1.3000 x 10-2, P0U = 1.3011 x 10-2

- C0 uncertain due to small monkey sample sizes: Nasal C0 = 2.49 ± 0.23, C0L = 2.11

Bone Marrow C0 = 17.5 ± 1.31, C0L = 15.34

- T1/2 and Cxss uncertain due to small rat sample sizes

12

Top Down Uncertainties (Human Analysis)

NPC: - VERY small number of deaths: 2 UnExp, 7 Exp - coarsely stratified cumulative exposure metric - marginally significant trend due to excess in

highest exposure category (3 deaths) - non-monotonic dose-response

LEU: - small number of deaths: 7 UnExp, 116 Exp - coarsely stratified cumulative exposure metric - non-significant positive trend due largely to

~ 47% deficit in Unexposed group relativeto the Exposed groups

- no dose-response in Exposed groups 13

Generalizing to Other Chemicals

• Methanol (metabolized to formaldehyde)

• Acetaldehyde (N2-hydroxyethyl-dG adducts

• Vinyl Acetate (metabolized to acetaldehyde)

• Vinyl Chloride (metabolized to chloroethylene oxide, producing 1 oxoethyl and 3 exocyclic etheno adducts)

• Ethylene Oxide (4 hydroxy-ethyl adducts)

14

Some Criteria for Use in Risk Assessment• Specific target sites in humans (epidemiology studies)

• Valid biomarkers of target site exposure that areplausibly correlated with the apical endpoint

• High precision/accuracy measurements thatdistinguish between endogenous / exogenoussources at low exogenous exposure levels

• Use conservative assumptions to fill data gaps

• Use to “reality check” and, when appropriate, replace top down analyses

15

Advantages of the Bottom Up Approach• Uses background cancer risk in humans

• Uses background (endogenous) adduct concentrations in humans, if available, or short-term animal data and equivalence assumptions

• Conservative:Linear at low doses (consistent with additivity)

All background risk attributed to endogenous adductsProvides an upper bound on low-dose slope

• Produces a completely independent “reality” check on risk extrapolations from high-dose data

16

Acknowledgments

17

James SwenbergBen MoellerKun LuAnn MasonRobinan Gentry